Overview
Display settings control how an agent's case data appears in Reindeer case views. They are configured on the agent through the external API and are separate from the case state itself.
Use display settings when reviewers, operators, or customer teams need to scan important case fields without opening raw JSON.
Questions This Article Answers
- Which case fields should appear in case views?
- How should display settings reference case state?
- Which fields are safe to show?
- What should be checked before changing display settings?
Display Configuration Pattern
Today, display settings support an optional agent icon and case-view configuration.
{
"icon": "invoice",
"cases": {
"fields": [
{
"label": "Vendor",
"path": "$.vendor.name"
},
{
"label": "Total",
"path": "$.invoice.total_amount"
},
{
"label": "Status",
"path": "$.validation.status"
}
]
}
}
The exact cases shape may vary by enabled UI view. Keep the configuration small and focused on fields that help users scan, filter, or decide what to do next.
Field Selection Guide
| Field type | Good fit for display? | Notes |
|---|---|---|
| Status or outcome | Yes | Useful for triage and review queues. |
| Customer, vendor, or account name | Yes | Keep names short and consistently formatted. |
| Document type | Yes | Good for filtering and routing. |
| Amount, date, or identifier | Yes | Use normalized values where possible. |
| Confidence score | Sometimes | Useful when reviewers need to understand automation risk. |
| Raw model reasoning | No | Too verbose and may change between runs. |
| Large nested objects | No | Better shown in detail views or omitted from display settings. |
| Sensitive data | Usually no | Prefer masked or derived values. |
Choose Stable State Paths
Display settings should reference fields that are stable across runs. Prefer paths that map to the case state schema, such as:
$.classification.type$.customer.name$.invoice.invoice_number$.invoice.total_amount$.validation.outcome$.review.required
Avoid pointing case views at temporary fields, raw model reasoning, large nested blobs, or values that may not be returned consistently by the external API.
Sensitive Data
Do not surface sensitive data in display settings unless your access model is designed for it. Prefer masked, summarized, or derived fields for personally identifiable information, credentials, financial details, or private customer content.
Examples:
- Show
masked_account_numberinstead of a full account number. - Show
review_required: trueinstead of the full reason if the reason may include sensitive text. - Show
document_typeinstead of the raw document body.
What the Configuring Team Provides
- The case state schema or sample case state.
- The fields that should appear in the case list or detail view.
- Labels for each field.
- Empty-state behavior for missing or null values.
- Any masking or redaction requirements.
- Review from the team responsible for access control and data handling.
Change Management
When you update the case state schema, review display settings at the same time. A display path that no longer exists may render as blank or unavailable in the UI.
Before enabling a display change, confirm:
- Each displayed path exists in the current case state schema.
- Test cases cover common, empty, null, and missing values.
- Long values remain readable.
- Sensitive fields are not exposed unintentionally.
- External integrations that depend on the same paths are still aligned.
Comments
0 comments
Please sign in to leave a comment.